Routing low-speed traffic requests onto high-speed lightpaths by using a multiobjective firefly algorithm

Abstract

Nowadays, the bandwidth requirements of the majority of traffic connection requests are in the range of Mbps. However, in optical networks each physical link is able to operate in the range of Gbps causing a huge waste of bandwidth as a result. Fortunately, using access station at each node of the optical network, several low-speed traffic requests may be multiplexed onto one high-speed channel. Multiplexing or grooming these low-speed requests is known in the literature as the Traffic Grooming problem - an NP-hard problem. Therefore, in this paper we propose the use of Evolutionary Computation for solving this telecommunication problem. The selected algorithm is an approach inspired by the flash pattern and characteristics of fireflies, the Firefly Algorithm (FA), but adapted to the multiobjective domain (MO-FA). After performing several experiments and comparing the results obtained by the MO-FA with those obtained by other approaches published in the literature, we can conclude that it is a good approach for solving this problem. © Springer-Verlag Berlin Heidelberg 2013.